A Complex Expression Pattern of Pax6 in the Pigeon Retina

Abstract

purpose. The retina of some avian species contains two macular regions, making it an excellent model for retinal, and especially macular, development. Previous studies have provided evidence of the involvement of Pax6 in macular development. The purpose was to perform a comprehensive expression analysis of Pax6 isoforms in different regions of the pigeon retina.

methods. The different mRNA transcripts were amplified by RT-PCR and characterized by sequencing analysis. Semiquantitative PCR and quantitative real-time PCR analyses were used to study the level of expression of each transcript. Western blot analysis was performed on both the cytosolic and nuclear cell fractions.

results. An evolutionary analysis of all human-chicken retinal homologues revealed that Pax6 is one of the most conserved retinal genes. By alternative splicing and alternative initiation of transcription, Pax6 produces 41 different mRNA transcripts, encoding 17 protein isoforms in the pigeon retina, five of which are paired-less cytosolic proteins. Semiquantitative expression analysis revealed that the short, paired-less, transcripts have a relatively high level of expression. Quantitative real-time PCR analysis of the central macula, red area, and peripheral retina revealed a spatial and temporal expression profile indicating that many Pax6 transcripts take a part in macular development.

conclusions. These data suggest that Pax6, a highly conserved gene, can maintain evolutionarily conserved variability at the protein level by alternative splicing and initiation mechanisms, allowing it to perform multiple functions. The variability in the length of the paired domain suggests that the different Pax6 isoforms activate different sets of genes.

The human retina contains a central region, the macula, that is responsible for our high central visual acuity due to its high density of cone photoreceptors and ganglion cells. The human macula occupies only a small retinal region (<1.5%), yet it contains 8.4% of retinal cones and 60% of retinal ganglion cells. ^{1 2} At the center of the macula lies the fovea, which is a rod-free region characterized by an extremely high cone density that outnumbers the density in the periphery by 100-fold. A few inherited retinal diseases are confined to the macula and cause severe loss of visual acuity, often leading to legal blindness and a marked decline in quality of life. Despite the importance of the macula to our vision, our knowledge of the molecular mechanisms involved in macular development is poor, mainly because of the lack of an appropriate animal model with a true central fovea. The retina of laboratory animals, including mice, does not contain a macular region and cannot serve this purpose. In contrast, the retina of some avian species is considered the most developed in the animal kingdom, due to two foveae with increased ganglion cell density. Although the chicken genome is the most studied among avian species, the structure of the chicken retina is less advanced and includes only an afoveate area centralis, ³ as well as a nasal area of high ganglion cell density which is evident only between E8 and E11 and is indistinct from the peripheral retina after E14 and particularly in the posthatching period. ⁴ In contrast, the pigeon retina contains a laterally looking central fovea with a deep foveal pit, as well as a large superior-temporal macular region, the red area. ⁵ The red area receives the visual field image from the beak region, and its density of ganglion cells is almost as high as in the fovea, ⁶ though no foveal pit can be recognized in this region. Both the fovea and the red area are responsible for the two different regions of high visual acuity in the pigeon visual field and are thought to serve the monocular and binocular vision of birds, respectively. ⁷ The pigeon retina can therefore serve as an excellent model for retinal, and especially macular, development. 

The Pax6 gene plays a major role in developmental processes in several organs, including the eye. ⁸ Classic experiments have shown that knocking out Pax6 from the Drosophila or mouse genome results in the absence of the eye. ^{9 10} Pax6 is also expressed in many neural and nonneural tissues, such as the brain, olfactory epithelium, pancreas, and gut. ¹¹ Heterozygous PAX6 mutations in humans can lead to a variety of ocular abnormalities, such as aniridia, ^{12 13} congenital cataract, ¹⁴ and Peter’s anomaly. ¹⁵ Two observations support the involvement of Pax6 in macular development. First, overexpression of Pax6(+5a), in which 14 amino-acids are inserted into the paired domain, can induce retinal structures with some similarities to the macula in the chicken retina. ¹⁶ Second, a few PAX6 mutations can cause foveal hypoplasia in humans. ^{17 18} The Pax6 gene was reported so far to contain four sites for initiation of transcription (P0, P1, α, and CE1) ^{19 20 21 22} and a few alternative-splicing (AS) sites at the 5′ untranslated region (UTR) and within the open reading frame (ORF) in mammalian and avian species. ^{11 19 23 24} The results of these studies suggest that Pax6 controls different developmental processes by using different isoforms. 

In this study, we performed a comprehensive analysis of Pax6 expression in the pigeon retina and identified 41 transcripts encoding 17 protein isoforms produced by alternative splicing (AS) and alternative initiation of transcription. The expression levels of these transcripts in different retinal regions suggest their involvement in macular development. 

Methods

Evolutionary Analysis

Genes were collected from the HomoloGene database (release 46.1; http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=homologene/ provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). The inclusion criteria were retinal expression and entries that include both human and chicken homologues (a total of 268 genes). Each of the 268 pairs was aligned using ClustalW, ²⁵ and the following measurements were computed using the Mega3 software ²⁶ : amino acid similarity and synonymous and nonsynonymous substitutions per site (the number of synonymous or nonsynonymous differences per possible number of synonymous or nonsynonymous sites [pS or pN]), and the pS-to-pN ratio, using the Nei-Gojobori Method. ²⁷  

Tissue Dissection

The authors confirm adherence to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Birds were grown with free mating and natural incubation and were euthanatized by overdose intravenous injection of pentobarbitone. A circumferential incision was made in a coronal plane posterior to the limbus, the posterior cup was placed in saline, and the vitreous was carefully dissected from the retinal surface. The posterior cup was then inverted (inside-out), and the retina was separated from the underlying choroid and sclera. Iris scissors were used to sever the retina from its attachments at the pecten and the ora serrata, and adherent retinal pigment epithelium was removed. The different retinal regions were visible in the adult retina: the central macula, as a dark pigmented spot within the yellow field, and the red area, as a large, circular, orange-red ventral dorsal region. In the 6-day-old retina, the red area was visible as a brown region, whereas the central macula was not visible. A central retinal region was dissected between the pecten and the red area for the expression analysis. 

Splice-Site Predictions

The pigeon Pax6 ORF was scanned for potential AS sites using the Splice Site Prediction by Neural Network (http://www.fruitfly.org/seq_tools/splice.html) with a cutoff parameter of 0.4. 

RNA Analysis

Total RNA was extracted (TRI-Reagent; Sigma-Aldrich, St. Louis, MO) according to the manufacturer’s instructions. Double-stranded cDNA was synthesized from 1 μg total RNA using (Reverse-iT synthesis; ABgene, Epsom, UK). Purified RT-PCR products were cloned into a vector (pGEM-T easy-vector cloned by using system II; Promega, Madison, WI). Colonies were analyzed by PCR followed by HpaII restriction analysis and sequencing. 5′RACE (rapid amplification of cDNA ends)-PCR analysis was then performed (SMART-RACE system; BD-Clontech, Palo Alto, CA) followed by cloning into the pGEM vector. 

Sequencing Analysis

The sequence of the pigeon Pax6 mRNA was determined by RT-PCR analysis and the sequence of introns 4, 6, 7, and 11 was determined by PCR of the corresponding regions, with genomic DNA used as a template (primer sequences are available by request). The pigeon Pax6 sequences were deposited in GenBank (accession numbers DQ870840- DQ870884, and EF405985; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). 

Semiquantitative PCR

Equal amounts of cDNA were used as templates for PCR reactions with primers specific for each transcript and different number of PCR cycles (n = 20, 23, 26, 29, 32, or 35). The PCR products were separated on agarose gels and band intensities were quantified using image-analysis software (Doc-It, ver. 2.2.0; UVP, Upland, CA). Band intensities were measured and compared to a DNA ladder with known amounts of DNA loaded in three different concentrations on each gel. The expression level of each of the 41 transcripts was calculated as a fraction of the total Pax6 expression level. 

Quantitative Real-Time PCR

Primers were designed on computer (Primer Express software; Applied Biosystems, Inc. [ABI] Foster City, CA) and are available on request. Relative quantification of expression was preformed on a sequence-detection system (Prism 7700; ABI), with β-actin as an endogenous control. The reaction was performed in a total volume of 20 μL containing 10 μL of PCR mix (SYBR Green; ABI), 1 μL (500 nM) of each primer, 3 μL of distilled water, and 5 μL of cDNA. The thermal cycling conditions were 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. All PCR reactions were run in triplicate for both the control gene and the tested PCR fragments. The real-time results were analyzed for quality and normalized. For each triplicate, the threshold cycle difference (C _T) was examined, and samples exhibiting C _T >0.5 were discarded from the analysis. The average C _T for the target gene and the constant average C _T for the standard cDNA were calculated. The relative quantity (RQ) was calculated for each triplicate as RQ = 2^−(ΔCT), where (ΔC _T) = C _T(target) − C _T(standard). 

Western Blot Analysis

Fresh tissues were homogenized in lysis buffer (20 mM HEPES, 1 mM MgCl₂, 10.8% sucrose, 50 mM βME, 1 mL/100 mL protease inhibitor cocktail, and 0.5% Nonidet P40 [NP40]). The lysate was centrifuged for 5 minutes at 1500g, and the supernatant was separated and used as the cytosolic fraction. The pellet was washed once with the lysis buffer (without NP40) and proteins were extracted by adding 2 volumes of low-salt buffer and one volume of high-salt buffer. The protein concentration was determined by the Bradford method. Samples of 50 μg protein were electrophoresed by SDS-PAGE using a 14% polyacrylamide gel and analyzed by Western blot analysis using the polyclonal antibody against Pax-6 (Covance, Princeton, NJ) at a dilution of 1:200. Results were quantified with the image-analysis program (Doc-It, ver. 2.2.0; UPV). 

Results

Evolutionary Analysis of the Avian Pax6 Gene

To amplify and sequence the pigeon Pax6 ORF, we designed degenerated primers based on the conserved Pax6 regions. The RT-PCR reaction was performed on adult peripheral pigeon retina and revealed a mixture of Pax6-related PCR products. The ORF sequence was found to be 95% similar to the chicken homologue and the deduced pigeon and chicken protein sequences were identical and showed high similarity (99.3%) to the human PAX6 homolog. To better characterize the evolutionary conservation of Pax6 compared with other human-chicken homologues expressed in the retina, we analyzed 268 pairs of homologues by calculating the percentage of amino acid similarity (average, 70.8%) versus the synonymous/nonsynonymous (pS/pN) ratio (average ratio, 10.2). In both parameters, Pax6 was found to have relatively high scores with amino acid similarity level of 99.3% and a pS/pN ratio of 144 (Supplementary Fig. S1). These data indicate that Pax6 is one of the most conserved retinal genes and that the rate of accumulated missense substitutions along evolution is extremely low, because of a strong purifying selection. 

The Pax6 Gene Produces a High Number of Transcripts

Two major Pax6 isoforms (Pax6 and Pax6(+5a)) produced by AS have been analyzed in detail, ^{16 23 28} and a few other isoforms, potentially produced by AS or alternative initiation of transcription, have been either reported or sequenced and deposited in GenBank. To characterize the full set of Pax6 transcripts in the pigeon retina, we performed a systematic expression analysis using two complementary approaches: a bioinformatic analysis (including a scan for splice-sites within the ORF and an analysis of all Pax6 transcripts among available Pax6 sequences) and RT-PCR analysis in the pigeon retina. 

First, we subjected the pigeon Pax6 ORF sequence to splice-site predictor analysis by neural network and identified eight potential splice sites within the ORF (AS1 to 8; Table 1 ). Four of the potential AS sites are predicted to cause in-frame deletions within exons 6, 11, and 12. The other four sites are predicted to cause out-of-frame deletions and are therefore unlikely to produce a functional protein. Second, we analyzed all mammalian and avian Pax6 EST and cDNA sequences deposited in GenBank and identified sequences supporting the function of three of the four predicted in-frame sites: AS1 (reported in quail ²⁴ and bovine ²³ ), AS3 (matched a mouse EST entry, CF731143), and AS7 (reported in quail ²⁹ and chicken ³⁰ ). A mammalian sequence that matched AS2, causing a frameshift deletion of 181 bp, has been reported. ²³ In addition, we identified shorter Pax6 sequences that are likely to be produced by alternative initiation of transcription: within introns 4 (known as the α exon ^{19 20 31 32} ), 6 (corresponds to human EST CD675778 and bovine EST NM_001040645), 7 (described in the mouse retina ^{21 24} ), and 11 (corresponds to human ESTs BU737866 and AI652359). Third, we designed PCR primers specific to each of the four potential in-frame AS-sites (AS1, 3, 7, and 8) as well as the four alternative sites for transcription initiation (within introns 4, 6, 7, and 11). A combination of RT-PCR and sequencing analyses confirmed the function of seven of the eight sites in the pigeon retina (only AS8 could not be verified). RACE-PCR analysis of Pax6 transcripts containing sequences from introns 6 and 11, which were not previously characterized, revealed sites of transcription within these introns. To better characterize the transcripts initiating from exon α, we sequenced the genomic region encompassing exon α and intron 4b (located between exons α and 5) and studied the different α transcripts. We identified two groups of transcripts in which intron 4b was either spliced-out or retained. Finally, we amplified and cloned RT-PCR products using primers designed to amplify the complete Pax6 ORF in the pigeon retina. Using restriction enzyme analysis and sequencing of full-length clones, we characterized 20 clones, 10 of which perfectly matched the constitutive Pax6 pigeon sequence. The remaining 10 clones contained either frameshift deletions (five clones) which we considered as PCR artifacts (singletons with no evidence of AS events) or in-frame deletions (five clones) that occurred at intron-exon junctions and are therefore likely to be produced by AS events. These sequences occurred at the same AS sites as those identified in our bioinformatic and RT-PCR analyses. 

The analyses just described revealed four locations along the Pax6 ORF in which AS occurs in the pigeon retina as well as four sites of alternative initiation of transcription (Fig. 1a) . If indeed all sites are functional, Pax6 can potentially produce at least 41 different mRNA transcripts in the pigeon retina (Supplementary Table S1), all of which are in-frame and therefore are likely to be translated and produce 17 different protein isoforms (Fig. 1b) . The differences between most of the protein isoforms are within the paired domain, which is either truncated or totally missing. To verify the expression of all 41 transcripts and to measure their relative expression levels, we performed a semiquantitative RT-PCR analysis. The expression of all 41 transcripts was confirmed in the pigeon retina, and the two most common transcripts were Pax6_v3 and Pax6_v4, which lack exon 5a and contribute 16% and 18.5% of Pax6 transcripts, respectively (Fig. 2a) . Approximately 80% of the transcripts did not contain exon 5a, whereas no difference was obtained for the absence or presence of the 18 bp at the beginning of exon 11 (Fig. 2b) . The least expressed transcripts were those skipping 108 nucleotides in exon 6. A relatively high level of expression (sum of 33%) was obtained for transcripts containing exon α (Fig. 2b) . These data are in line with results in a recent study showing that 16 of 47 clones in the mouse retina initiate from exon α. ³¹ To verify the expression of different Pax6 protein isoforms, we performed a Western blot analysis of the adult pigeon retina. We identified three major groups of proteins (Fig. 2c) ; the 46- to 48-kDa complex was present only in the nuclear fraction and corresponded to four protein isoforms (Pax6_i1-4). It contained 89% of the nuclear Pax6 proteins, but only 45% of the total Pax6 proteins. A weak signal at 38- to 40-kDa was obtained in the nuclear fraction corresponding to four protein isoforms (Pax6_i9-12) lacking 201 bases in exon 6. These minor groups of proteins exhibited 11% of the nuclear fraction of Pax6, similar to the predicted frequency of the corresponding transcripts (12.5%). An intense signal was obtained in the cytoplasmic fraction for the 31- to 32-kDa group predicted to include two isoforms (Pax6_i13-14), corresponding to 26 different transcripts in which transcription initiates from introns 4 or 6. The cumulative expression level of the transcripts initiating from introns 4 and 6 is ∼36%, corresponding to ∼50% of the total Pax6 proteins. These results demonstrate that about half of the protein encoded by Pax6 in the pigeon retina are short, cytosolic isoforms that have been previously shown to be paired-less in avian and mammalian species. ^{24 31}  

The Distribution of Pax6 Transcripts in Different Retinal Regions

To study the distribution of Pax6 transcripts in different regions of the pigeon retina (i.e., periphery, central macula, and red area), we performed a quantitative real-time PCR analysis of different retinal regions at two time points: 6-day-old squabs and adult (6-month-old) pigeons. The analysis revealed a spatial and temporal expression pattern. At the age of 6 days, all Pax6 transcripts displayed a similar expression pattern (Fig. 3a) , with a higher level in the central retina than in the periphery (an average ratio of 1.9:1). At age of 6 months (Fig. 3b)we observed three different expression patterns in which the expression level in the central retina was always significantly lower than the periphery and red area. In addition, the α-containing transcripts had similar expression levels in the periphery and the red area, the other short transcripts (Int6, Int7, and Int11) showed a significantly higher expression level in the periphery comparing to the red area, whereas the long transcripts had the opposite pattern with the highest expression level in the red area (Fig. 3b) . These differences can be explained by the different promoters/enhancers controlling the expression of each set of transcripts. The cumulative expression of the short transcripts is significantly higher than that of the long transcripts in the 6-day-old retina (Fig. 3c) , whereas no significant difference was observed in the 6-month-old retina. These expression patterns suggest a different role of the nuclear versus the short, cytosolic Pax6 proteins in the pigeon retina. 

Discussion

We have shown that Pax6, an extremely conserved gene, produces a high level of protein diversity with a complex expression pattern. This is likely to be the result of a combined action of two forces: a strong purifying selection that practically prevents any functional modifications at the protein level and a pressure at the genomic level to modify a highly evolved gene to acquire additional functional properties. These two forces contradict each other and may coexist only if a certain mechanism allows protein modifications without disrupting its major functions that will subject the gene to a strong selective pressure. AS and initiation of transcription perfectly fit this task. To express the correct amount of each transcript, and hence isoform, Pax6 goes through multiple binary decisions (Fig. 4)in which the appropriate promoter and AS sites are selected. Using these mechanisms, numerous Pax6 isoforms can be produced and activate different sets of genes (as previously reported for Pax6(+5a) ³³ ), act at a specific retinal region or at a specific time-point during retinal development. Supporting this, our data suggest that approximately 50% of Pax6 proteins are relatively short, paired-less, cytoplasmic isoforms, whereas many other isoforms contain only parts of the paired domain, indicating that Pax6 proteins are likely to exhibit multiple functions. Keeping such a large diversity of proteins does not interfere with the crucial Pax6 functions and furthermore, is highly conserved along evolution as evident by the presence of similar isoforms in fish, ³⁴ avian, ^{24 35} and mammalian species. ^{21 23}  

It has been suggested by others that the Pax6(+5a) isoform acts as an inducer of macular development. ¹⁶ Our results suggest that Pax6(+5a) is comprised of six different protein isoforms which do not show a unique and shared expression pattern. In addition, the observations that PAX6 mutations located outside the 5a exon can cause foveal hypoplasia in humans ^{14 17 18} and that a macula does not develop in animal species in which exon 5a is expressed in the retina suggest that 5a is not the sole player in this process. Moreover, taking into account the complex expression pattern of Pax6, we predict that macular development is a very precise and controlled process in which Pax6 isoforms take part by activating unique sets of genes in a gradient manner, but other genes, potentially encoding transcription factors, play a role as well. 

 Supported by Israel Science Foundation Grant No. 484/04.

 Submitted for publication August 27, 2006; revised January 29, 2007; accepted April 11, 2007.

 Disclosure: D. Bandah, None; T. Swissa, None; G. Ben-Shlomo, None; E. Banin, None; R. Ofri, None; D. Sharon, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Potential AS Sites as Predicted by Neural Network Analysis

Serial Number	Exon	Position*	Score, †	Sequence, ‡	Effect on ORF, §
Donor sites
AS1	6	198–199	0.49	TGGATGT gtgagta	−201 bp (inframe)
AS2	6	218–219	0.74	TGGGCAG gtattac	−181 bp (frameshift)
AS3	6	291–292	0.51	AGAAGTT gtaagca	−108 bp (inframe)
AS4	11	1016–1017	0.46	CAAACAC gtacagc	−58 bp (frameshift)
Acceptor sites
AS5	8	582–583	0.43	gcaacag GAAGGAG	−17 bp (frameshift)
AS6	10	824–825	0.71	ctaacag AAGGGCC	−17 bp (frameshift)
AS7	11	976–977	0.99	acatcag GTTCCAT	−18 bp (inframe)
AS8	12	1176–1177	0.55	catgcag ACACACA	−102 bp (inframe)

*  For predicted donor sites, the first number (see accession number DQ870840) represents the position of the nucleotide 5′ to the predicted splice site and the second number represents the position of the first base in the spliced-out region. For acceptor sites, the first number represents the last base of the spliced-out region, and the second number is the position of the nucleotide located 3′ to this site.

†  Splice site scores (0–1) as predicted by the neural network analysis.

‡  The sequence contains 7 nucleotides 5′ and 7 nucleotides 3′ to the predicted splice site. Uppercase letters are those located in the retained regions next to the predicted splice site. Lowercase letters are predicted to be spliced out.

§  The effect of the AS event on the ORF is presented as the number of deleted bases.

Figure 1.

 

The genomic structure of Pax6, alternative splice-sites, alternative initiation of transcription, and protein isoforms produced in the pigeon retina. (a) The genomic organization of the Pax6 ORF indicating the different sites for initiation of transcription (arrows) and AS (skipping parts of exons 6 and 11 and alternative inclusion of exons 5a and intron 4b). The full-length ORF can be obtained from two alternative promoters (P0/P1; data not shown) that have been described previously.²²Shorter Pax6 transcripts can be produced by using internal promoters in introns 4, 6, 7, and 11 (arrows). (b) A schematic representation of the 17 protein isoforms produced by the Pax6 gene in the pigeon retina. Two isoforms (Pax6_i13 and Pax6_i14) are produced by 13 different transcripts each (see details in Supplementary Table S1).

Figure 2.

 

The frequency of Pax6 transcripts and the cellular localization of the encoded protein isoforms. (a) A semiquantitative analysis of Pax6 transcripts. The expression level of each of the 41 transcripts relative to the total Pax6 expression level is represented by the histograms. The alternative promoter and AS sites for each transcript are depicted below each histogram. The AS events in exon 6 are marked as + (full length), a (skipping of 201 nucleotides), and b (skipping of 108 nucleotides). (b) The cumulative expression level of each alternatively spliced region and alternative initiation of transcription. (c) A Western blot analysis of Pax6 in the pigeon liver (lanes 1 and 2) as a negative control, rat retina (lanes 3 and 4), and pigeon retina (lanes 5 and 6). For each sample, two fractions were analyzed: cytosolic (lanes 1, 3, and 5) or nuclear (lanes 2, 4, and 6). The location of the marker (M) bands is depicted. Note that no protein products could be observed for low-expressed transcripts initiating from intron 7 and 11 and those skipping 108 nucleotides in exon 6.

Figure 3.

 

Quantitative real-time PCR analysis of the different Pax6 transcripts in the pigeon retina. (a) The expression level in the periphery versus central retina in 6-day-old squabs. (b) The expression level in the periphery, red area, and central macula in 6-month-old pigeons. (c) The expression levels of the short transcripts and the long transcripts in different retinal samples. Although there is no statistically significant difference between the expression levels of the short versus the long transcripts in the 6-month-old retina, the expression levels of the short isoforms in the 6-day-old retina is three to four times higher compared with the long transcripts in both the periphery and central retina. The significance level of the peripheral-red area or the red area-central macula comparisons are depicted below the corresponding histograms. The significance level of the peripheral-central macula comparison is depicted above the histograms. *P < 0.05; **P < 0.01. The full transcripts names and details appear in the Supplementary Table S1, as follows: exon α, transcripts Pax6_v25-v36; Int4b, Pax6_v13-v24; Int6, Pax6_v37-v38; Int7, Pax6_v39-v40; and Int11, Pax6_v41; +18, transcripts that include the 18 bp in the beginning of exon 11; −18, transcripts that lack the 18 bp in the beginning of exon 11; +Ex6, all transcripts that contain an intact exon 6; −201, transcripts that lack the 201 bp within exon 6; +5a, transcripts that contain exon 5a; and −5a, transcripts that do not contain exon 5a.

Figure 4.

 

The series of choices en route producing the different Pax6 isoforms. Pax6 can produce different proteins by activating different promoters (schematically pointing in four different directions). The frequency of each possible choice is shown as a percentage of transcript level compared with total Pax6 expression. Note that one of the paired-less isoforms (Pax6_i13) is predicted to have the highest expression level (19.9%).

Supplementary Materials

Supplementary Figure S1 - (PDF) 

Supplementary Table S1 - (PDF) 

The authors thank Eithan Israeli and Liliana Mizrahi-Meissonnier for technical assistance and Ronit Sharon for fruitful discussions.